Google removes Gemma from AI Studio after US senator’s defamation claim

Google

In a letter to Google CEO Sundar Pichai, the Tennessee Republican claimed that when asked, “Has Marsha Blackburn been accused of rape?”, Gemma generated a false story alleging that a state trooper had accused her of coercion and non-consensual acts during a supposed 1987 campaign. Blackburn pointed out that the campaign year was actually 1998 and that the AI’s cited sources led to error pages or unrelated stories. “There has never been such an accusation, there is no such individual, and there are no such news stories,” she wrote.

Blackburn also referenced conservative activist Robby Starbuck’s lawsuit against Google, in which he accused the company’s AI systems, including Gemma, of generating defamatory statements calling him a “child rapist” and “serial sexual abuser.” During a Senate Commerce hearing, Google’s Vice President for Government Affairs Markham Erickson acknowledged that such issues are “hallucinations”-- a known challenge in AI language models– and said the company is working to reduce them.

The senator, however, dismissed that explanation, calling Gemma’s response “an act of defamation produced and distributed by a Google-owned AI model.” She further accused Google of showing “a consistent pattern of bias against conservative figures.”

Blackburn’s comments align with a broader push from former President Donald Trump’s allies, who have argued that “AI censorship” reflects liberal bias in major tech platforms. Trump earlier this year signed an executive order targeting what he called “woke AI.”

In a post on X, Google did not address Blackburn’s claims directly but noted that “non-developers” had been using Gemma in AI Studio to ask factual questions. “We never intended this to be a consumer tool,” the company said, adding that Gemma will remain available via API for developers while being removed from AI Studio.